{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/adagan-boosting-generative-models","title":"AdaGAN: Boosting Generative Models","arxiv_id":"1701.02386","date":"2017-01-09","proceeding":"NeurIPS 2017 12","authors":["Ilya Tolstikhin","Sylvain Gelly","Olivier Bousquet","Carl-Johann Simon-Gabriel","Bernhard Schölkopf"],"abstract":"Generative Adversarial Networks (GAN) (Goodfellow et al., 2014) are an\neffective method for training generative models of complex data such as natural\nimages. However, they are notoriously hard to train and can suffer from the\nproblem of missing modes where the model is not able to produce examples in\ncertain regions of the space. We propose an iterative procedure, called AdaGAN,\nwhere at every step we add a new component into a mixture model by running a\nGAN algorithm on a reweighted sample. This is inspired by boosting algorithms,\nwhere many potentially weak individual predictors are greedily aggregated to\nform a strong composite predictor. We prove that such an incremental procedure\nleads to convergence to the true distribution in a finite number of steps if\neach step is optimal, and convergence at an exponential rate otherwise. We also\nillustrate experimentally that this procedure addresses the problem of missing\nmodes.","url_abs":"http://arxiv.org/abs/1701.02386v2","url_pdf":"http://arxiv.org/pdf/1701.02386v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"adagan-boosting-generative-models","repo_url":"https://github.com/tolstikhin/adagan","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"BSD-3-Clause"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1701.02386","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}